Also, I think there should be a significant fee for asset issuances, otherwise we'll get spammed with "parked" assets from people who just want to reserve nice names.
The problem with having fees for asset issuances is finding a way to keep them floating. Any fixed fee we would pick now would be too high, and then too low, if Counterparty scales as I hope it does. If anyone has a suggestion for a fee schedule that will scale well, I'd love to hear it.
The current system does allow for the transference of asset ownership, so valuable names may indeed be sold for their market value when appropriate.
I think speculation is the solution to that problem. Channel speculation so that the fee always rests in perfect equilibrium with the market. The speculation will find the equilibrium point.
You could use a sort of distributed betting mechanism to allow players to predict what they expect the fees to be (guess the price of the fee for x?). So if Alice predicts the fee should be over 20 XCP and Bob predicts it should be under 20 XCP they should both bet and put their XCP on the line. The fee should be high enough that neither of them will want to be wrong. One of them will have to be right. So you can actually use speculation to find the price.
Bitshares is going to use speculation to determine the price of bitUSD. The price of Bitshares is determined by speculation via Protoshares. To determine the price for issuing an asset we have to speculate. Speculation will produce a price discovery where the market will decide the maximum amount it's willing to pay to issue an asset. That maximum amount should be the current fee. This would mean the fee would have to be dynamically adjusted in such a way that it adapts to speculation somehow.
One way I can think of to do this is to somehow have a decentralized data feed which is based on the outcome from betting/speculation. That data feed becomes the XCP fee price chart or XCP fee price index. Once the market or the network has made it's decision (consensus) the fees should be set for that hour, or that day.
I don't have it completely figured out though. I may have figured out that speculation can do it in theory but I don't have a detailed practical explanation for how to do it. I just believe that is ultimately how the problem will be solved and the direction you should take.